Table 5.

CNN model training results.

	Dataset	Without Sampling	CallBacks	Random under Sampler	SMOTE	SMOTE Tomek	Borderline SMOTE	ADASYN
Accuracy	CTU13	0.997520	0.998044	0.971556	0.997805	0.995140	0.993538	0.993538
	IoT23	0.952365	0.945287	0.892822	0.896551	0.945287	0.892836	0.896751
Precision	CTU13	0.886515	0.868871	0.195595	0.761460	0.588351	0.701167	0.517816
	IoT23	0.736959	0.845727	0.995560	0.995560	0.997780	0.999970	0.999989
Recall	CTU13	0.736959	0.845727	0.995560	0.995560	0.997780	1	1
	IoT23	0.991621	1	0.886644	0.890587	1	0.886660	0.890785
F-Score	CTU13	0.804848	0.857143	0.326955	0.862915	0.740222	0.824337	0.682317
	IoT23	0.975221	0.971874	0.939904	0.942116	0.971874	0.939912	0.942233
FPR	CTU13	0.000659	0.000892	0.028612	0.002179	0.004879	0.002978	0.006507
	IoT23	0.725878	1	0.000437	0.000402	1	0.000455	0.000175